KernelBench-Mega
PulseAugur coverage of KernelBench-Mega — every cluster mentioning KernelBench-Mega across labs, papers, and developer communities, ranked by signal.
-
Fable 5 tops KernelBench-Mega, outperforming Claude Opus and GPT 5.5
Fable 5 has achieved the top position on the KernelBench-Mega benchmark, demonstrating a significant advancement in AI's capability for R&D tasks. The AI model generated Cuda code that resulted in an 18.71X speedup on a…
-
AI model Fable writes faster GPU kernels and automates online work · 2 sources tracked
A new AI model named Fable has demonstrated the ability to write efficient GPU kernels, achieving an 18.71X speedup on an Nvidia RTX Pro 6000 Blackwell Workstation Edition compared to an optimized PyTorch baseline. This…
-
Claude Opus 4.8 leads KernelBench-Mega benchmark, outperforming NVIDIA GPUs
A new benchmark called KernelBench-Mega has been released, which involves rewriting GPU megakernels for each generated token. The benchmark was tested on NVIDIA's RTX PRO 6000, H100, and B200 GPUs, with Claude Opus 4.8 …